The Science of Mindfulness (part 2)

This is the second part of my article on the science of mindfulness. The first part can be found here. Both are a follow-up on my initial post, where I discussed a Belgian TV documentary (rather well made) about the "hype" of mindfulness in Belgium.

In the comments on this post, I got rightly called out on my first assessment of the (lack of) scientific proof for mindfulness. I'm especially grateful for the comments made by Shirley Rieven. Most of what follows is based on the articles she gathered, although any idiocy in interpreting them remains mine, of course...

While in the first part focus was on specific treatment in rather specific cases (avoiding relapse after depression), this part will focus more on the different (potential) domains where mindfulness can be applied. In that sense, it is also more inline with the TV documentary, where it was shown that, apart from depression, it could also be applied to stressful managers and kids with Attention Deficit Disorder (ADD). The latter seems already rejected by the researchers quoted in the previous part, but reduction of stress in general does seem to be a topic where a lot of research is going on.

A lot of people base their opinion on the meta-analysis performed by Grossmann and colleagues, calledMindfulness-based stress reduction and health benefits. It appeared in 2004 in theJournal of Psychosomatic Research, and a full copy of the text can be found via Google Scholar or here. The study, according to Google Scholar, is referenced more than 1300 times in other scientific articles.

Basically, 64 studies (published and unpublished) were selected that claimed health benefits for a variety of problems (e.g. chronic pain) or for people under stressful conditions (e.g. medical students) when applying Mindfulness-Based Stress Reduction (MBSR). Studies were selected inasmuch as they adhered to a certain standard 8-10 week group program, with exercises as also detailed in my previous post. Of course, quantitative outcomes had to be available, based on a standardised scaling.

Interestingly, controlled studies needed to have a control group (no intervention or controlling for the social aspects of the MBSR group), but one study was also thrown out of the data aggregation because it was a "comparison" study that compared its effectiveness to a well established psychotherapy. Now that was of course a study that interested me. Although it is unpublished, I managed to track down the article (not peer reviewed) here on the servers of the University of Nijmegen in The Netherlands. It claims that mindfulness has indeed better outcomes when compared to a standard procedure. However, the study is based on a low number of participants (25), and it seems that Major Depressive or manic bipolar patients were also not included in the experiment. As noted in the first part of my article, the more seriously mentally ill patients are probably not the target for mindfulness therapy. Also, something that was rather odd was that two authors (including the leading) seemed to be from the faculty of science, mathematics and computing science, and one from the social sciences. That seems odd indeed for a psychotherapy article, but I will not speculate on what it means for the content of the article.

Anyway, back to the metastudy. The above article was only one of the 44 articles removed from final consideration. I think it both speaks for the thoroughness of Grossman to be very strict, and in my opinion the still weak state that mindfulness research was before 2003. 20 out of 64 is not a good statistic. Most exclusions, according to Grossman, were due to not strictly following mindfulness as laid out in the initial definition, or because insufficent statistical details were available.

In the end however, the statistic is even worse. Only ten studies were finally included in the meta-analysis, for reasons mentioned above. The study mentions the effect sizes of these studies, which are small. Of the total 771 subjects participating, the biggest study was on 277 students (stress reduction). You also have to consider that half of those 771 subjects got the mindfulness treatment, the other ones the control.

My point is not that this is bad. Given the results of the meta-analysis, it seems indeed that there is an effect. The mean effect size is just above .5, indicating a rather large positive impact of MBSR. But given the few studies included, and given that each of them has a rather low sample size, I'm still not entirely convinced. Also, and Grossman is entirely correct in pointing this out, is that the studies and the meta-study have measured the immediate effect. There was insufficient data available for long-term follow-up, so this means that more research is needed to check if it isn't some temporary effect that is in the long run indistinguishable from other treatments.

Of course, this study was done in 2003, so maybe in the mean time there were some evolutions. That will be my topic for the third part in this series, with an overall conclusion. Maybe the follow-up studies have been done in the intervening 10 years, and maybe more studies are now available to check if the initial result holds.

Part three is this way.